Eye state detection method and computer readable storage medium

Abstract

The invention discloses an eye state detection method and a computer readable storage medium. The method comprises the steps of obtaining a to-be-detected image and obtaining an eye region image; performing binarization and normalizationon the eye region image; traversing the eye region image, and sequentially obtaining a pixel point as a current pixel point; according to a preset feature extraction operator, extracting a feature value corresponding to the current pixel point, wherein the feature extraction operator is a square mask image, the pixel value of the pixel point of a mask area in the mask image is 1, pixel values of the pixel points of other areas are 0, and the mask area is centered and is in a preset centrosymmetric geometrical shape; generating an eye feature vector corresponding to the to-be-detected image according to the feature value corresponding to each pixel point in the eye region image; and inputting the eye feature vector into the trained classifier to obtain an eye state corresponding to the to-be-detected image. According to the invention, the eye state can be accurately identified in the low-resolution image.

Description

technical field [0001] The present invention relates to the technical field of image processing, in particular to an eye state detection method and a computer-readable storage medium. Background technique [0002] With the rapid development of image processing technology, facial feature point calibration algorithms have achieved good results, which can accurately calibrate the contours of key organs such as eyes, nose, and mouth, such as dlib, openpose, etc. Blink detection can be well used in human-computer interaction, such as unlocking mobile phones after face recognition, blinking to control robots, etc. Ideally, it becomes very simple to identify the state of the eye (blink detection) based on this type of algorithm, as long as the eye contour is used to easily identify the state of the eye. [0003] However, the calculation complexity of the openpose algorithm is very high. When the current high-performance discrete graphics card 1080ti runs this algorithm, it can onl...

AI Technical Summary

Problems solved by technology

[0003] However, the calculation complexity of the openpose algorithm is very high. When the current high-performance discrete graphics card 1080ti runs this algorithm, it can only process more than a dozen images per second, not to mention the small smart terminal with relatively low computing power, which is difficult to achieve smooth to run

Although the facial feature point calibration of the dlib algorithm runs very fast and has good results in the daily use of mobile phone software, but when the camera is far away from the person, the eye resolution is very low, and the calibrated eye contour is not enough Accurate, miscalibration often occurs, resulting in blink recognition errors, which is not conducive to remote human-computer interaction

Images